A Technological Infrastructure for Implementing a Policy of Condition Based Maintenance for a Fleet of Railway Vehicles

Different policies exist to maintain a fleet of railway vehicles or machineries in general, e.g. cyclic, event driven, proactive, condition based. Each policy offers a different level of performances, and it requires a different amount of data for its actuation. As a general rule, the higher the diagnostic information, the more effective are the maintenance actions. Effectiveness of maintenance can be measured in terms of availability of each vehicle (or machine), and must be compared with its cost. Increasing the effectiveness of maintenance implies creating a network for funneling information from the machineries to the central maintenance room where information is elaborated. Moreover, to increase the amount of information the machineries must be sensorised to monitor variables of interest for detecting incipient failures or need of repairing interventions. Such an infrastructure of sensors, data collection units, communication networks, and computers for data storing and sorting has a cost and a failure rate. Increasing the power of this infrastructure increases its cost and its failure rate, so it is mandatory for a Company to find an adequate equilibrium point between complexity and expected benefits. This paper presents a roadmap for a Company that manages a fleet of vehicles (or machineries) towards the Condition Based Maintenance, that is the most efficient maintenance policy. The paper associates to each maintenance policy its benefits and its requirements in terms of infrastructure and costs. Using the proposed roadmap, a Company can decide its strategy and can schedule the investments in terms of technologies and skills. Bombardier Transportation Italy started this roadmap few years ago, for moving from a reactive maintenance policy to a proactive policy where some indicators are measured and monitored to optimize (and reduce) the maintenance interventions. The paper presents the first results achieved by means of an analytical tool developed together with the Dept. DITEN of the University of Genoa applied to the fleet of Bombardier locomotives in Europe. Even very simple statistical analysis on the data collected from the fleet of vehicles gave quite interesting and economically important results.